American Hockey League attendance: a study of fan preferences for fighting, team performance, and promotions.

Paul, Rodney J. ; Weinbach, Andrew P. ; Robbins, Daniel 等

American Hockey League Attendance: A Study of Fan Preferences for Fighting, Team Performance, and Promotions

The hockey world has recently been met with unfortunate tragedies. Deaths of current and past players, who typically played the role of enforcers on their teams, have again turned the focus of debate on to the role of fighting in professional hockey. Enforcers are players who earn their living by protecting other members of their team through their physical play. This physical play often manifests itself as on-ice fighting with opposing team players. Discussions of the role of fighting and its impact on brain injuries has led to renewed calls for a ban on fighting in the sport of hockey.

Wade Belak, Derek Boogaard, and Rick Rypien, who all recently played in the NHL, the top hockey league in the world, each passed away in 2011. Although none of these deaths were directly related to on-ice fights during a game, medical experts and many in the media have suggested that fighting contributed to the early deaths of these players. In addition to these young players, retired longtime NHL enforcer Bob Probert also died of a heart attack in 2011 at the age of 45. Probert was a very popular enforcer during his career in the NHL. In autopsy, it was found that a degenerative disease known as chronic traumatic encephalopathy existed in the brain of Bob Probert. This disease has been linked to the brains of boxers and other deceased athletes in highly physical sports, including another former NHL player (and known fighter), Reggie Fleming. It was speculated that an important factor leading to this degenerative disease in the brain of Probert was a long history of blows to the head. This again increased the fervor of those opposed to fighting in the NHL to call for its ban.

On the medical side, the debate about fighting has recently been waged by experts in the field. Dr. Rejendra Kale (2011), in an editorial in the Canadian Medical Association Journal, called for an outright ban of fighting in hockey due to its impact on brain injuries. Similar sentiments were raised by a neurosurgeon in Toronto, Dr. Charles Tator (2009), who stated that clear evidence exists that blows to the head lead to degeneration of the brain in hockey players. Dr. Robert Cantu, an expert in head trauma, also echoed these sentiments.

On the other side of the battle of medical experts were Dr. Ruben Echemendia and Dr. David Milzman. Dr. Echemendia (2001), former president of the National Academy of Neuropsychology, stated there was not enough evidence to establish a link between hockey fights and brain injuries. Dr. Milzman, of the Georgetown University School of Medicine, stated in an interview with HealthDay News that his working research revealed that fights in hockey rarely cause any injuries, including brain trauma.

As this debate continues to rage in the media, the medical profession, and throughout the hockey world, the obvious question that needs to be asked is why would professional hockey leagues keep fighting in the game if there is any evidence that fighting may lead to brain injuries. One answer lies in the game itself, as many players, coaches, and analysts state that the game is actually safer due to fighting, as the players police themselves. Without fighting in the game, they argue that dangerous stick work and hits will increase if there is not fear of retribution in the minds of the players on the ice. Therefore, they argue, the enforcer plays a pivotal role on the ice as it keeps the game and its players in line.

Another reason that fighting is likely kept in the game has nothing to do with the play on the ice, but is due to simple economics. Fans appear to enjoy fighting and are willing to pay to watch hockey games which include fighting. If fighting is removed from the game, there is the possibility that fewer fans will attend games and/or be willing to pay less to watch the on-ice action.

Given the importance that fighting may play in the consumption of hockey tickets, the role of violence (on-ice fights) in the NHL has been studied previously in Jones (1984), Jones, Ferguson, and Stewart (1993), Jones, Stewart, and Sunderman (1996), and Paul (2003). Fighting was shown to have a positive and significant effect on attendance in these studies. Although fighting is not allowed in the DEL league in Germany, evidence was found that physical play (penalty minutes) increased attendance (Coates et al., 2011). The same result related to physical play was not found to affect attendance in the SM-Liiga in Finland (Coates et al., 2011). Fighting was not shown to have an impact on attendance in junior hockey in the Quebec Major Junior Hockey League (Paul & Weinbach, 2011).

Fighting and physical play may play a role in attendance in some leagues, but it does not appear to influence the win percentage of teams. If win percentage, in turn, impacts willingness on the part of fans to purchase tickets, this could also impact the overall profitability of the league and its teams. Leard and Doyle (2011) studied fighting success, determining winners and losers of individual fights in the NHL, and concluded that winning fights does not affect the probability of a team winning a game. Coates et al. (2011) found a negative relationship between team success and fighting and penalty minutes in the NHL.

With the renewed interest in the role that fighting plays in the sport of hockey, we believe it is important to understand the impact of fighting on attendance in other professional leagues, namely the top developmental league for the NHL, the American Hockey League (AHL). The AHL is the AAA-equivalent (in baseball language) level of play for professional hockey players. Although many European leagues exist, where players may hone their skills in hopes of landing a job in the NHL, and some players may directly jump to the NHL from juniors or college, the AHL is typically thought of as the main stepping stone on the path to the top-level of professional hockey. A full list of AHL teams and nicknames is provided in Appendix 1 of this paper.

This paper investigates the role of fighting as it relates to AHL attendance through a regression model. Fights are noted for each game of the season for each team and are tabulated as the season progressed, creating a running average of fights-per-game, which is included as an independent variable within the regression equation. In addition to the role of fighting, the impact of team performance (winning and scoring), parent team success, distance to the parent team city, city demographics, and typical factors such as the day of the week and the month of the season are analyzed.

Another unique attribute which we investigate in this paper as it relates to attendance in the AHL is the role of promotions. We were able to gather each of the promotions (merchandise giveaways, group nights, concerts, etc.) for all of the AHL teams for the 2010-11 season. We have included this information as a series of dummy variables within the regression equation to determine which of these promotions have a statistically significant impact on attendance and to examine the relative effectiveness of each promotion in terms of the additional fans they attract.

American Hockey League Attendance Model

The goal of this study is to identify the impact of fighting, winning, and promotions on attendance in the AHL. Previous literature surrounding hockey attendance has mainly focused on the effects of fighting and rule changes. Past studies of the National Hockey league include Jones (1984), Jones, Ferguson, and Stewart (1993), Jones, Stewart, and Sunderman (1996), and Paul (2003). Attendance at minor league games have been studied recently by Hong (2009) and Rascher, Brown, Nagel, and McEvoy (2009).

Our regression model is simple and straightforward, using per-game attendance as the dependent variable. We use a reduced-form model as ticket price is not included in the model. Given that teams set prices at the beginning of the season, presumably to maximize expected profits, the price does not change for game tickets over the course of the season for each team. The inclusion of average ticket price is problematic as well, as different levels and prices of tickets exist for each team. Therefore, we assume that the profit-maximizing price set by the teams will be captured through other demographic variables of the city.

Independent variables are grouped by categories which include timing of the game (days of the week dummies--Wednesday omitted dummy, month of the year January omitted dummy), promotions, demographics (Population and income per capita), affiliate NHL team success (points earned in previous season), distance from NHL team (in miles), and on-ice team performance (Win Percentage, Total Goals Scored per Game, Fights per game average). Days of the week and months of the year dummy variables are included in the regression to account for daily and monthly effects. Weekends and months later in the season (during the playoff push) are expected to show positive and significant results. Wednesday is the omitted dummy for the days of the week and January is the omitted dummy for the months of the hockey season, with all other daily and monthly results compared to these omitted categories.

The on-ice performance variables were broken into the effects of winning (win percentage), scoring, and fighting. The AHL uses a point-based standings system, with two points for a win, one point for an overtime loss or shoot-out loss, and no points for a regulation loss. Therefore, win percentage was calculated by the number of points attained by the home team out of the total points possible (two times the number of games played). This variable was calculated as a running average and the value at any given time is the percentage of possible points achieved going into the current home game. If fans value a winning team at the minor league level, instead of only valuing marketing gimmicks or seeing certain players who are approaching the major league level, the win percentage entering the game is expected to have a positive and significant effect on attendance.

Scoring is also calculated as a running average going in the current home game. If fans prefer to see higher-scoring games, this variable should have a positive and significant effect on attendance. Scoring likely serves as a proxy for overall game excitement, which we would expect fans to prefer. Scoring average is included for both the home and visiting teams in the regression model.

Fighting in hockey is a hotbed issue for the media, fans, and the leagues in general. Past studies on the effects of fighting have shown that increases in fighting have led to increases in attendance. A positive and significant effect of fighting on attendance was shown for teams based in the United States in Jones, Stewart, and Sunderman (1996). Using data from a decade later, the 1999-2000 season, positive increases in attendance were found in relation to fighting for both U.S. and Canadian based teams (Paul, 2003).

The effect of violence in sports, particularly hockey, is of great interest to researchers in many disciplines. Determining the importance of fighting for hockey at the minor league level will allow a deeper exploration of this topic. Fighting was calculated on a per-game basis (as with the scoring and winning variables above) and the variable is the average fights per game going into the current home game.

Promotional data was taken from the team websites, which listed game promotions along with their season schedule. Promotions can have an important impact on per-game attendance for sports teams. Promotions may even be more important at the minor-league level, where team winning may not be the most important factor to fans when making the decision to purchase tickets. Promotions have been studied before in the literature, for sports such as Major League Baseball (McDonald & Rascher, 2000).

The promotional information available from the teams was made into a series of dummy variables representing many different possible categories. If the goal of these promotions is to bring a greater number of fans to the arena, their effects should be positive and significant. Due to the very small amount of sell-outs during the AHL season, effects of seats sold due to promotions being minimized due to sell-outs is not expected to be a problem.

The full list of promotional categories used as dummy variables in the regression model is given in Table 1 below. The first column lists the name of the promotional category, the second column gives a brief description of the promotions which fall into this category, and the third column gives the frequency of the listed promotion during the 2010-11 AHL Season (1200 total games).

Table 2 continues to present frequencies of dummy independent variables in the regression model. This table notes the frequency of games on each day of the week, each month of the season, and the number of early starts. Early starts are games beginning before the normal evening start time (around 7 p.m.). These early game starts include afternoon contests on weekends and games with morning starts, used as a promotional tool.

The regression model includes 1,200 observations from the 2010-11 AHL season. Given issues with heteroskedasticity, the regression was run with White's heteroskedasticity-consistent standard errors and co-variances. Those adjusted results are what are presented in the table below. Two separate columns show the results with different functional forms of the dependent variable. Specification I is has attendance in levels, specification II uses the natural log of attendance as the dependent variable with the independent variables in levels (log-lin), and specification III uses a log-log model with the dependent variable and non-binary independent variables in logs. Regression results using groups of similar promotions, to have fewer categories to compare and contrast results, are shown in Appendix 2. Statistical significance in the regression results is noted with ^-notation as * represents significance at the 10% level, ** at the 5% level, and *** at the 1% level.

Results in all three specifications (levels, log-lin, log-log), are similar in terms of impact and statistical significance. Overall, the results of both regressions appear to tell the same basic story. The results below are discussed mainly in terms of the results in levels (specification I), for ease of explanation, but differences between the regression results in terms of model specification are also noted in the discussion below.

In relation to city demographics, population was shown to have a positive and significant effect on attendance. This reveals that cities with larger populations attract more fans to AHL games. From the regression results, it appears that there is an increase in attendance of nearly one fan for each additional 1,000 people in the city. Income per capita was shown to have a positive and significant effect on attendance, implying AHL hockey as a normal good, in the first model specification (levels), but was shown to have an insignificant (yet positive) effect in the second model specification. From specification I, it appears that a $100 increase in income per capita leads to about three or four more fans in attendance in AHL cities. Income levels of the AHL city may play a role in attendance, but it does not appear to play as statistically significant a role as population.

In relation to the NHL parent club of the AHL team, the success of the NHL affiliate was shown to have a positive and significant effect on attendance at the AHL. The more successful the parent franchise, the higher the attendance level at its top developmental league city. This result could be due to fans taking a greater interest in the team when the parent team is performing well or it could denote an overall better quality of players (due to scouting, coaching, management, etc.) in some organizations compared to others. In any case, the parent team success seemed to filter to the minor league level. The distance, in miles, to the parent NHL club did not have a significant effect on attendance. In alternative specifications, where distance squared and cubed were included, statistically significant results were also not found in relation to this variable.

The monthly effects were mostly statistically insignificant, other than the impact of October, the first month of the AHL season. With the inclusion of an opening night dummy variable in the promotional dummy categories, the remaining October games are rather poorly attended across the cities of the AHL with over 1,600 fewer fans attending games near the start of the season. This result was found to be statistically significant at the 1% level.

The day of week dummy variables revealed expected results as weekend games are much more popular than mid-week games. Each weekend day, Friday, Saturday, and Sunday, was shown to have positive and significant results at the 1% level. This result likely stems from the opportunity cost of fans on weekends compared to weekdays. Weekend games offer a greater chance for many people to attend, given fewer time commitments, such as work and school. Saturday was the most popular day, with an increase in attendance of nearly 2,300 compared to the omitted day, Wednesday. Sunday attracted over 1,500 more fans, while Friday night games attracted over 1,400 fans, compared to Wednesday. Early starts were shown to have a negative, but insignificant impact on game attendance.

In relation to the on-ice performance variables, attributes of the home team played an extremely important role. Team success, measured in relation to winning games (based on total points earned out of total possible points--to account for overtime and shootout games as well as those ended in regulation time), was shown to have a positive, but statistically insignificant effect on attendance. Fans were shown to respond to home team scoring in a positive and significant manner. Teams that scored more often, a proxy for more excitement during the game, were able to draw additional fans to games. For each additional goal scored, on the average, teams saw increases in attendance of over 770 fans.

The key variable in the study, the role of fighting, is shown to heavily influence attendance in a positive and significant manner. For each additional fight per game that the home team averages, attendance at AHL games increased by almost 1,000 fans. Perhaps a more realistic example of adding an additional fight every fourth game would still increase attendance by about 250 fans for each game. This result was found to be statistically significant at the 5% level. Fans are influenced by the number of fights that the home team participates in during the course of the season. The visiting team fights per game average was not found to be statistically significant. This result likely stems from an information asymmetry as it is somewhat easy (by attending games or reading the box score in the local paper or team internet site) to follow fighting frequency for the home team, but the transactions costs of following fighting frequency for the opponent are considerably higher. Despite the tragedies of hockey players during the 2010-11 season and the loud voices of the opponents to fighting in the game of hockey, hockey fans still responded quite favorably to witnessing a fight during a game.

In relation to the opponent, fans did respond to the quality of the opposing team, as opponent points per game average was shown to have a positive and significant effect on attendance. Fans of the AHL appear to prefer to see their home team play the best teams in the league. Division rivals, however, were not shown to have a significant impact on fan attendance as either the frequency of seeing these divisional opponents play or the wide range and large number of teams in the divisions did not allow for more heated rivalries and the likely increase in attendance that those rivalries would inspire.

The dummy variable categories related to promotions revealed some interesting results with important ramifications for team officials scheduling promotions for their teams. First, not all promotions were the same with respect to fan response. Many of the promotions were not found to have statistically significant effects on attendance. In addition, some of the promotions were actually found to have a negative impact on attendance for games. These insignificant or negative promotions could be a result of trial-and-error by teams, thereby learning that some promotions do not actually work. Other insignificant or negative promotions could be a result of sponsorship deals, where a firm will pay to have its name associated with a given promotion. If this sponsorship deal is sizeable enough, the team may still be better off having the promotion despite the fall in ticket sales for that game.

To begin with the promotions that were shown to have positive and significant effects on attendance, it is important to first note that some promotions had a much larger impact on attendance than others. The promotions that led to more than 1,000 additional fans were opening night, closing night, fan appreciation nights, post-game concerts, and valuable theme-park ticket giveaways. Opening night, closing night, and fan appreciation nights (which some teams run for a few games, some on the last game, some on the last few games) are typical fan-favorites and therefore the success of the promotions and events associated with these games and the overall pageantry of these games was not surprising. The high-end prime events (concerts and theme park pass giveaways) boosted attendance by over 4,000 fans each, but it should be noted these are rather expensive promotions and they occurred infrequently across the AHL schedule. As with any promotion, cost-effectiveness is a major issue for the hockey organization. Costs of promotions were not available for our analysis, but given that team management knows these costs, analysis of the increase revenues due to the promotion (increased ticket sales, parking concessions, food and drink, etc.) compared to the cost of the promotion would be a straightforward analysis to determine the economic viability of various promotions.

It should be noted that "bring your dog to the rink nights" did have a positive impact on the average, attracting over 1,600 additional fans, but the result was not found to be statistically significant in specification I (dependent variable in levels). However, in specification II (dependent variable in logs), the dog nights were shown to have a positive and statistically significant effect on attendance at the 10% level. From these results, we surmise it is likely that there is great variation across the league in that fans in some cities really enjoy bringing their dog to the rink, while fans in other cities do not.

Merchandise giveaways, special jerseys, and college nights each attracted over 700 additional fans for these games. It appears AHL hockey fans enjoy team merchandise giveaways and respond favorably to their team wearing a unique jersey such as pink jerseys for breast cancer awareness or throwback jerseys. College students also appear to enjoy the discount provided to them on "college nights" and respond quite favorably to these promotions.

Two of the promotions with negative and significant impacts on attendance were those related to discounted tickets for the game. Although lower prices would be expected to attract a greater number of fans, this result is likely more of a function of the discounted price being a signal to fans that the games are appearing on weeknights or against unpopular opponents. The systematic scheduling by some teams in the AHL (having consistent discounts on Tuesdays or Thursdays), is likely to have split the negative impact of weekday games between the day-of-week dummy variables and these promotional categories. Both the general discounted ticket promotion and the "Kraft Singles" nights (where Kraft singles wrappers could be exchanged for discounted tickets to the game) were shown to have a negative impact on attendance.

The other negative and significant promotion was part of a series of promotions aimed at children in attendance. Although it would be thought that something like this would be popular, where children could collect all of the promotions in a series to make a set, it appeared to actually backfire and led to fewer fans in attendance. This could be a result of parents fearing the disappointment of the child if they could not attend all of the game in the series to complete the set. Instead of attending these games with their children, it appeared many fans (over 1,000) simply did not attend these games. This may provide an important lesson for team management in the future as this result would not have been expected. In addition, other negative effects of promotions were found, based on model specification, for family nights and airline discount raffles.

Conclusions

AHL attendance for the 2010-11 season was studied using an extensive data set including all promotions for every team in the league. A simple regression model with per-game attendance as the dependent variable was specified and a wide range of independent variables were examined. The main goal of this research was to establish the link, or lack thereof, between fighting and attendance in the American Hockey League.

AHL fans were shown to be significantly impacted by fighting during hockey games as fights per game of the home team had a large positive and significant effect. Fan preference for fighting in the AHL supports results shown for the NHL in previous research (Jones, Stewart, & Sunderman, 1996; Paul, 2003). This result sheds some light on the controversy surrounding fighting in hockey, especially in light of the recent tragic deaths of NHL enforcers discussed in the introduction to this paper. It appears a key reason to keep fighting in the sport of hockey is that fans respond favorably at the gate to on-ice fights. This is an important consideration, as professional sports is a business, and needs to be carefully weighed against the actual risks of brain injuries to players as this debate continues into the future.

Although the role of fighting on hockey attendance may have been the key variable of interest, other findings from the regression model were also compelling. In relation to demographics, AHL attendance was shown to increase with the size of the population. Weekends were the most popular days for games and early season contests in October (other than opening night) were poor draws. With respect to on-ice performance, visiting team success (in terms of number of points achieved out of total points--akin to win percentage) was shown to have a positive and significant effect on attendance. Fans also were shown to prefer home teams which played in higher-scoring games, as this is likely a proxy for within-game excitement.

With respect to highly detailed promotional data for each team, promotions were generally shown to be popular with AHL fans as many types of promotions were shown to have positive and significant effects. The most popular promotions were shown to be post-game concerts, opening night, closing night, fan appreciation nights, and theme park passes. Other positive and significant results were shown for promotions involving merchandise giveaways, college nights, and special jersey nights.

Overall the AHL appears to be a model league in terms of the economic factors which impact attendance. Fans react as would be expected, with team quality, scoring, and fighting being key drivers of fans to purchase tickets. Promotions overall appeared to be a success, with teams trying more and more ways to attract fans to games of this top-level minor league. This research confirms many factors that sports economists believe impact attendance and put numerical estimates and statistical significance to a wide array of factors which influence attendance in the American Hockey League. In addition, it confirms what many observers of professional hockey already knew or suspected; fighting at hockey games influences the financial success of a team and league by significantly impacting ticket sales.

Appendix 1: AHL Teams and Nicknames 2010-11

Atlantic Division

Portland Pirates Manchester Monarchs Connecticut Whale Worcester Sharks Providence Bruins Springfield Falcons Bridgeport Sound Tigers

North Division

Hamilton Bulldogs Lake Erie Monsters Manitoba Moose Abbotsford Heat Toronto Marlies Grand Rapids Griffins Rochester Americans

East Division

WB/Scranton Penguins Hershey Bears Charlotte Checkers Norfolk Admirals Binghamton Senators Syracuse Crunch Adirondack Phantoms Albany Devils

West Division

Milwaukee Admirals Houston Aeros Peoria Rivermen Texas Stars Oklahoma City Barons Chicago Wolves San Antonio Rampage Rockford Icehogs

Appendix 2: Regression Results with Grouping of Promotional
Variables
                              Specification I:
                             Dependent Variable:   Specification II:
                             Attendance (Levels)     Log-Lin Model

Intercept                       -5873.22 ***          6.4306 ***
                                  (-3.5272)            (22.5213)
Population                         0.0009            0.0000001 ***
                                (2.9252) ***           (4.3108)
Income Per Capita                  0.0328              0.000003
                                  (1.5177)             (0.8045)
NHL Affiliate Points             41.1819 ***          0.0097 ***
                                  (4.3729)             (5.7210)
Distance to NHL Affiliate          -0.1341             -0.000003
                                  (-1.2879)            (-0.1567)
October                         -1698.11 ***          -0.3754 ***
                                  (-4.9985)            (-6.2816)
November                          32.2707               -0.0643
                                  (0.0660)             (-1.1589)
December                          -184.0747            -0.0959 *
                                  (-0.4308)            (-1.711)
February                          208.1375              0.0361
                                  (0.6776)             (0.6927)
March                             122.0272              0.0520
                                  (0.3830)             (0.9324)
April                             164.9150              0.0482
                                  (0.4456)             (0.7555)
Sunday                           1457.96 ***          0.2459 ***
                                  (2.6907)             (3.0636)
Monday                            308.0805              0.1306
                                  (0.7148)             (1.5745)
Tuesday                           -464.8779             -0.0113
                                  (-1.5247)            (-0.2191)
Thursday                          47.3514               0.1223
                                  (0.0878)             (1.4553)
Friday                          1457.653 ***          0.3532 ***
                                  (5.1749)             (7.9697)
Saturday                        2149.875 ***          0.4132 ***
                                  (5.1393)             (8.9169)
Early Start                      -806.5785 *            -0.0318
                                  (-1.6992)            (-0.5094)
Division Opponent                 -182.5432             0.0132
                                  (-0.8065)            (0.4944)
Home Team Points Per Game         712.0027              0.0383
                                  (1.0996)             (0.3813)
Home Team Scoring               871.1071 ***          0.1793 ***
  Average                         (2.8147)             (3.2894)
Home Team Fights Average         1163.61 ***           0.1826 **
                                  (2.6612)             (2.4036)
Visiting Team Points            921.4916 ***           0.1573 **
  Per Game                        (2.6040)             (2.3536)
Visiting Team Scoring             -192.5149            -0.0518 *
  Average                         (-1.0243)            (-1.7171)
Visiting Team                     445.4447              -0.0004
  Fights Average                  (0.8506)             (-0.0067)
Giveaways                       761.3039 ***          0.1447 ***
                                  (2.8051)             (4.8837)
In Arena Event                    187.4518             0.0384 *
                                  (0.9627)             (1.6851)
Discount                         -405.2710 *            -0.0618
                                  (-1.7212)            (-1.2248)
Group                             49.7840               -0.0343
                                  (0.1528)             (-0.9295)
Concessions                       96.6932               0.0255
                                  (0.6486)             (1.0699)
Closing Night/Fan App.           1150.14 ***          0.1896 ***
                                  (5.0506)             (5.7254)
Opening Night                    2271.37 ***          0.4834 ***
                                  (4.6530)             (6.9985)
R-squared                          0.1953               0.3714

                             Specification III:
                               Log-Lin Model

Intercept                        2.3701 ***
                                  (2.2609)
Population                       0.0709 ***
                                  (3.7331)
Income Per Capita                 0.0494
                                  (0.5704)
NHL Affiliate Points             0.8882 ***
                                  (6.1816)
Distance to NHL Affiliate         -0.0022
                                 (-0.1611)
October                         -0.3771 ***
                                 (-6.0589)
November                          -0.0637
                                 (-1.0595)
December                         -0.0965 *
                                 (-1.6808)
February                          0.0323
                                  (0.6128)
March                             0.0516
                                  (0.8953)
April                             0.0437
                                  (0.6918)
Sunday                           0.2461 ***
                                  (3.0914)
Monday                            0.1249
                                  (1.5082)
Tuesday                           -0.0221
                                 (-0.4202)
Thursday                          0.1207
                                  (1.4019)
Friday                           0.3492 ***
                                  (7.7744)
Saturday                         0.4228 ***
                                  (9.0288)
Early Start                       -0.0182
                                 (-0.2934)
Division Opponent                 0.0147
                                  (0.5465)
Home Team Points Per Game         0.0768
                                  (0.7176)
Home Team Scoring                0.5940 ***
  Average                         (3.9309)
Home Team Fights Average         0.1539 **
                                  (2.0292)
Visiting Team Points             0.1482 **
  Per Game                        (2.1129)
Visiting Team Scoring             -0.1300
  Average                        (-1.5174)
Visiting Team                     0.0022
  Fights Average                  (0.0422)
Giveaways                        0.1402 ***
                                  (4.7497)
In Arena Event                    0.0324
                                  (1.4091)
Discount                          -0.0730
                                 (-1.4110)
Group                             -0.0678
                                 (-1.5342)
Concessions                       0.0185
                                  (0.7516)
Closing Night/Fan App.           0.1860 ***
                                  (5.6137)
Opening Night                    0.4759 ***
                                  (6.9517)
R-squared                          0.3637

Description of Grouping of Promotional Variables:
Giveaways--Giveaway, Series Promotion, Series Promotion Youth,
Theme Park
Promotion
In Arena Event--Special Jersey, Post-Game Skate, Autograph,
Theme Night, Teddy
Bear Toss, Donation, Dog to Arena, Post-Game Concert,
In-Game Wedding, Airline
Discount Raffle
Discount--Discount Tickets, Free Youth, Free Future
Tickets, Singles Night
Group--Family Night, Local Fire and Police Department Night,
Boy Scout Night, College Night, Religious Group, Grandparents Night
Concessions--Food and Drink, Beer
Closing Night/Fan App.--Closing Night, Fan Appreciation Night
Opening Night--Opening Night

References

Coates, D., Battre, M., & Deutscher, C. (2012). Does violence in professional ice hockey pay? Cross country evidence from three leagues. Sports Economics: Management and Policy, 4(2), 47-63.

Echemendia, R., Putukian, M., Mackin, R., Julian, L., & Shoss, N. (2001). Neuropsychological test performance prior to and following sports-related mild traumatic brain injury. Clinical Journal of Sport Medicine, 11(1), 23-31.

Hong, J. (2009). A comparison of motivational factors affecting attendance between avid and casual fans at minor league hockey games. International Journal of Sports Management and Marketing, 5(1-2), 115-131.

Jones, J. C. H. (1984). Winners, losers, and hosers: Demand and survival in the National Hockey League. Atlantic Economic Journal, 12(3), 54-63.

Jones, J. C. H., Ferguson, D. G., & Stewart, K. G. (1993). Blood sports and cherry pie: Some economics of violence in the National Hockey League. American Journal of Economics and Sociology, 52(1), 87-101.

Jones, J. C. H., Stewart, K. G., & Sunderman, R. (1996). From the arena into the streets: Hockey violence, economic incentives, and public policy. American Journal of Economics and Sociology, 55(2), 231-49.

Kale, R. (2011). Stop the violence and play hockey. Canadian Medical Association Journal, 184(3), 275.

Leard, B., & Doyle, J. (2011). The effect of home advantage, momentum, and fighting on winning in the National Hockey League. Journal of Sports Economics, 12(5), 538-560.

McDonald, M., & Rascher, D. (2000). Does bat day make cents? The effect of promotions on the demand for Major League Baseball. Journal of Sport Management, 14, 8-27.

Paul, R. J. (2003). Variations in NHL attendance: The impact of violence, scoring, and regional rivalries. American Journal of Economics and Sociology, 62(2), 345-364.

Paul, R. J., & Weinbach, A. P. (2011). Determinants of attendance in the Quebec Major Junior Hockey League. Atlantic Economic Journal, 39(3), 303-311.

Rascher, D., Brown, M., Nagel, M., & McEvoy, C. (2009). Where did National Hockey League fans go during the 2004-05 lockout? An analysis of economic competition between leagues. International Journal of Sport Management and Marketing, 5(1-2), 183-195.

Tator, C. (2009). Concussions are brain injuries and should be taken seriously. The Canadian Journal of Neurological Sciences, 36(3), 269-270.

Rodney J. Paul [1], Andrew P. Weinbach [2], and Daniel Robbins [3]

[1] St. Bonaventure University

[2] Coastal Carolina University

[3] Florida Atlantic University

Rodney J. Paul is a professor in the School of Business. His research interests include sports gambling markets and fan demand for sporting events.

Andrew P. Weinbach is an associate professor in the Department of Accounting, Finance and Economics. His research interests include identifying factors that influence attendance and television ratings and market efficiency in sports wagering markets.

Daniel Robbins is a graduate student. His research interests include studying fan demand for hockey at the major and minor league levels.

Table 1: Description and Frequency of Promotions in the AHL
2010-11 (1200 Games)

Promotion             Description                             Frequency

Giveaway              Promotional giveaway                       247
Special Jersey        Team wears special jersey--auctioned      55
                      off after game (usually for charity)
Discount Tickets      Discount tickets offered to game           58
Free Youth Tickets    14 and under get in free (with adult)       5
Free Future Tickets   Tickets to future games are offered        32
Family Nights         Package of 4 tickets, food, soda, and      104
                      popcorn or program
Food/Beverage         Discounted food or beverage                110
                      (non-alcohol)
Beer                  Discount beer                              84
Post-Game Skate       Fans skate on ice after game               50
Autograph             Pre- or post-game autograph session        57
Theme Night           Games such as Halloween games, Pink        137
                      in the Rink, Seinfeld Night, other
                      single-game themes
Teddy Bear Toss       Bring Teddy Bear to be game--thrown       22
                      on ice at end to benefit charity
Fan Appreciation      Fan appreciation game--usually with       23
                      many festivities
Series Giveaway       Series of giveaways                        38
Police/Fire           AHL game and local police                   8
                      department/fire
Department/School     department or high school/college
                      game
Game
Boy Scout Day         Boy Scout themed games                     17
College Discount      Discount ticket with college ID            38
Donation Night        Some form of donation being accepted       46
                      at the arena
Bring your Dog        Bring your dog to the game                  7
Religious             Religious theme night                       6
Theme Park            Theme park pass giveaway                    3
Singles Night         Theme night for singles--discount         11
                      tickets
Grandparents Free     Grandparents get in free                    2
Postgame Concert      Concert following game                      8
On-Ice Wedding        Wedding on ice during game (3 teams)        3
Airline $500          $500 airline gift card given away           3
Opening Night         First game--typically with                30
                      festivities
Closing Night         Last game--typically with                 30
                      festivities

Table 2: Frequency Tabulations for Days of Week, Months of Season,
and Early Starts--AHL 2010-11 (1200 Games)

Category       Frequency

Sunday            227
Monday            24
Tuesday           96
Wednesday         147
Thursday          31
Friday            331
Saturday          344
October           145
November          188
December          186
January           207
February          180
March             206
April             88
Early Start       297

Summary statistics for attendance, the dependent variable in
the regression model, and the non-binary independent variables
are given in Table 3.

Table 3: Summary Statistics--AHL 2010-11 Season

Variable     Attendance   Population   Income   Scoring
                                        Per       Per
                                       Capita    Game
                                                (Team)

Mean          5,484.53     529,265     24,746    2.83

Standard      3,815.44     760,603     7,409     0.38
Deviation

Median         4,813       177,165     23,823    2.79

Variable     Fights      NHL      Distance
              Per     Affiliate    to NHL
              Game     Points       Team

Mean          0.98      92.28      809.03

Standard      0.23      13.34      885.36
Deviation

Median        0.97       96         554

Table 4: Regression Results--AHL Attendance 2010-11 Season

                             Specification I:
                            Dependent Variable:   Specification II:
                            Attendance (Levels)     Log-Lin Model

Intercept                      -6034.41 ***          6.4160 ***
                                 (-3.6855)            (23.4240)
Population                      0.0009 ***          0.0000001 ***
                                 (2.9136)             (4.2051)
Income Per Capita                0.0364 *             0.000003
                                 (1.7244)             (1.1258)
NHL Affiliate Points            41.4186 ***          0.0096 ***
                                 (4.6150)             (6.0422)
Distance to NHL Affiliate         -0.0753             0.00001
                                 (-0.6262)            (0.8782)
October                        -1669.27 ***          -0.3603 ***
                                 (-5.0243)            (-6.1646)
November                         60.9504               -0.0678
                                 (0.1241)             (-1.2670)
December                         -99.0493              -0.0766
                                 (-0.2294)            (-1.4003)
February                         284.6235              0.0439
                                 (0.9651)             (0.8889)
March                            143.7303              0.0561
                                 (0.4728)             (1.0535)
April                            229.5211              0.0567
                                 (0.6831)             (0.9612)
Sunday                         1536.400 ***          0.2579 ***
                                 (2.7616)             (3.2306)
Monday                           365.4994              0.1361
                                 (0.8432)             (1.5594)
Tuesday                          -327.5932             0.0095
                                 (-1.0690)            (0.1814)
Thursday                         35.4050               0.1183
                                 (0.0672)             (1.4832)
Friday                          1457.34 ***          0.3522 ***
                                 (5.0621)             (7.9643)
Saturday                       2298.532 ***          0.4303 ***
                                 (5.1209)             (9.1201)
Early Start                      -786.5852             -0.0114
                                 (-1.5588)            (-0.1800)
Division Opponent                -175.4578             0.0143
                                 (-0.6989)            (0.5444)
Home Team Points                 887.0521              0.0749
Per Game                         (1.3591)             (0.7778)
Home Team Scoring               770.4126 **          0.1474 ***
Average                          (2.5343)             (2.7567)
Home Team Fights                994.0925 **           0.1626 **
Average                          (2.3890)             (2.2072)
Visiting Team Points            870.1583 **           0.1551 **
Per Game                         (2.4383)             (2.3028)
Visiting Team                    -143.8547             -0.0402
Scoring Average                  (-0.7680)            (-1.3978)
Visiting Team Fights             489.5016              -0.0042
Average                          (0.3651)             (-0.0752)
Giveaway Item                  993.9418 ***          0.1713 ***
                                 (3.1182)             (5.1149)
Special Jersey                  722.8293 *           0.1532 ***
                                 (1.7793)             (3.0849)
Discounted Tickets             -755.9858 **          -0.1559 **
                                 (-2.4584)            (-2.1789)
Free Youth Tickets               272.5138              0.0935
                                 (0.3983)             (0.8305)
Free Future Tickets              361.2629              0.1141
                                 (0.7237)             (1.3160)
Family Nights                    -323.1577           -0.1346 ***
                                 (-0.4328)            (-2.5788)
Discount Food                    284.3888              -0.0218
and/or Beverage                  (0.8993)             (-0.3925)
Discount Beer                    -255.2176             0.0579
                                 (-0.7803)            (1.0988)
Post-Game Skate                  109.7331              0.0077
                                 (0.2920)             (0.1345)
Autograph Night                   -163.65              -0.0064
                                 (-0.4884)            (-0.1180)
Theme Night                      -41.1968              -0.0116
                                 (-0.1482)            (-0.2770)
Teddy Bear Toss                  -637.1049             -0.1217
                                 (-1.1013)            (-0.9860)
Fan Appreciation Night          1027.690 **          0.1834 ***
                                 (2.1653)             (2.6023)
Series Promotion                 -82.9877              0.0838
                                 (-0.2377)            (1.5817)
Series Promotion Youth         -1021.041 **            -0.1145
                                 (-1.9636)            (-1.0704)
Local Police/Fire                -456.5098             0.0384
Department/                      (-0.6532)            (0.5067)
School Game Night
Boy Scout Night                  198.6178              0.0741
                                 (0.3282)             (0.9771)
College Night                   982.6358 **           0.1585 **
                                 (1.9645)             (2.0428)
Donation Night                   153.4786              0.0968
                                 (0.2717)             (1.2759)
Bring Dog to the Stadium         1627.836             0.2997 *
                                 (1.3961)             (1.6689)
Religious Theme                  -509.7306             0.0498
                                 (-0.8409)            (0.5805)
Theme Park Promotion            4562.58 ***          0.6558 ***
                                 (5.9631)             (5.0617)
Singles Night                   -1955.75 **          -0.1513 **
                                 (-2.4576)            (-2.0950)
Grandparents Night               -197.6183             -0.6980
                                 (-0.1077)            (-0.8326)
Post-Game Concert               4728.67 ***          0.7525 ***
                                 (7.6288)             (13.7181)
In-Game Wedding                  1389.05               0.2763
                                 (1.4405)             (1.4221)
Airline Discount Raffle         -1005.78 *             -0.0699
                                 (-1.8771)            (-1.0780)
Opening Night                   2113.36 ***          0.4549 ***
                                 (4.2886)             (6.3023)
Closing Night                   1122.09 ***          0.1739 ***
                                 (2.9253)             (2.7213)
R-squared                         0.2152               0.4103

                            Specification III:
                              Log-Log Model

Intercept                       2.0851 **
                                 (2.0755)
Population                      0.0679 ***
                                 (3.6807)
Income Per Capita                0.0838
                                 (0.9587)
NHL Affiliate Points            0.8794 ***
                                 (6.4578)
Distance to NHL Affiliate        0.0073
                                 (0.5736)
October                        -0.3597 ***
                                (-5.9388)
November                         -0.0652
                                (-1.1125)
December                         -0.0766
                                (-1.3643)
February                         0.0413
                                 (0.8263)
March                            0.0573
                                 (1.0272)
April                            0.0556
                                 (0.9413)
Sunday                          0.2500 ***
                                 (3.1766)
Monday                           0.1287
                                 (1.4883)
Tuesday                          -0.0059
                                (-0.1108)
Thursday                         0.1147
                                 (1.4035)
Friday                          0.3454 ***
                                 (7.6762)
Saturday                        0.4367 ***
                                 (9.1716)
Early Start                      0.0104
                                 (0.1689)
Division Opponent                0.0179
                                 (0.6755)
Home Team Points                 -0.0307
Per Game                        s(-0.3010)
Home Team Scoring               0.4931 ***
Average                          (3.2065)
Home Team Fights                0.1429 **
Average                          (1.9645)
Visiting Team Points            0.1450 **
Per Game                         (2.0327)
Visiting Team                    -0.0960
Scoring Average                 (-1.1716)
Visiting Team Fights             -0.0040
Average                         (-0.0786)
Giveaway Item                   0.1644 ***
                                 (4.9500)
Special Jersey                  0.1603 ***
                                 (3.0878)
Discounted Tickets              -0.1698 **
                                (-2.3355)
Free Youth Tickets               0.1060
                                 (0.9448)
Free Future Tickets              0.0752
                                 (0.8795)
Family Nights                  -0.1669 ***
                                (-3.2932)
Discount Food                    -0.0002
and/or Beverage                 (-0.0032)
Discount Beer                    0.0165
                                 (0.3049)
Post-Game Skate                  -0.0232
                                (-0.4066)
Autograph Night                  -0.0222
                                (-0.3734)
Theme Night                      -0.0107
                                (-0.2501)
Teddy Bear Toss                  -0.1158
                                (-0.9361)
Fan Appreciation Night          0.1888 ***
                                 (2.7297)
Series Promotion                 0.0888
                                 (1.5284)
Series Promotion Youth           -0.1171
                                (-1.0433)
Local Police/Fire                0.0416
Department/                      (0.5346)
School Game Night
Boy Scout Night                  0.0667
                                 (0.8283)
College Night                    0.1282 *
                                 (1.6364)
Donation Night                   0.1053
                                 (1.4289)
Bring Dog to the Stadium         0.2812
                                 (1.5458)
Religious Theme                  0.0165
                                 (0.1891)
Theme Park Promotion            0.7444 ***
                                 (4.6150)
Singles Night                    -0.0942
                                (-1.4111)
Grandparents Night               -0.7083
                                (-0.8437)
Post-Game Concert               0.6957 ***
                                (11.9489)
In-Game Wedding                  0.2390
                                 (1.3063)
Airline Discount Raffle         -0.1213 *
                                (-1.7384)
Opening Night                   0.4471 ***
                                 (6.2326)
Closing Night                   0.1603 **
                                 (2.4806)
R-squared                        0.4007